JOURNAL OF NEUROTRAUMA 32:13121323 (September 1, 2015)

Mary Ann Liebert, Inc.

DOI: 10.1089/neu.2014.3644

Early Surgery versus Initial Conservative Treatment

in Patients with Traumatic Intracerebral
Hemorrhage (STITCH[Trauma]):
The First Randomized Trial
Downloaded from online.liebertpub.com by 36.80.211.182 on 09/24/17. For personal use only.

A. David Mendelow,1 Barbara A. Gregson,1 Elise N. Rowan,1 Richard Francis,1 Elaine McColl,2
Paul McNamee,3 Iain R. Chambers,4 Andreas Unterberg,5 Dwayne Boyers,3
and Patrick M. Mitchell 6 on behalf of the STITCH(Trauma) Investigators

Abstract
Journal of Neurotrauma 2015.32:1312-1323.

Intraparenchymal hemorrhages occur in a proportion of severe traumatic brain injury TBI patients, but the role of surgery
in their treatment is unclear. This international multi-center, patient-randomized, parallel-group trial compared early
surgery (hematoma evacuation within 12 h of randomization) with initial conservative treatment (subsequent evacuation
allowed if deemed necessary). Patients were randomized using an independent randomization service within 48 h of TBI.
Patients were eligible if they had no more than two intraparenchymal hemorrhages of 10 mL or more and did not have an
extradural or subdural hematoma that required surgery. The primary outcome measure was the traditional dichotomous
split of the Glasgow Outcome Scale obtained by postal questionnaires sent directly to patients at 6 months. The trial was
halted early by the UK funding agency (NIHR HTA) for failure to recruit sufficient patients from the UK (trial regis-
tration: ISRCTN19321911). A total of 170 patients were randomized from 31 of 59 registered centers worldwide. Of 82
patients randomized to early surgery with complete follow-up, 30 (37%) had an unfavorable outcome. Of 85 patients
randomized to initial conservative treatment with complete follow-up, 40 (47%) had an unfavorable outcome (odds ratio,
0.65; 95% confidence interval, CI 0.35, 1.21; p = 0.17), with an absolute benefit of 10.5% (CI, - 4.425.3%). There were
significantly more deaths in the first 6 months in the initial conservative treatment group (33% vs. 15%; p = 0.006). The
10.5% absolute benefit with early surgery was consistent with the initial power calculation. However, with the low sample
size resulting from the premature termination, we cannot exclude the possibility that this could be a chance finding. A
further trial is required urgently to assess whether this encouraging signal can be confirmed.

Key words: craniotomy; intracerebral hemorrhage; randomized; controlled trial; traumatic brain injury

Introduction
I n the UK, there are 1.4 million presentations of traumatic brain
injury (TBI) at emergency departments each year.1 The inci-
dence worldwide varies between 56 and 430 per 100,000 popula-
tion each year,2 with the highest incidence being in Asia (344 per
100,000) and the lowest in the United States (103 per 100,000).3
The mortality rate for severe isolated TBI in the UK varied between
16% and 40%,4 which is similar to the world-wide rates that vary
between 15% and 38%.3 Intracranial hemorrhage occurs in more
than 60% of serious TBIs in one or more of three types: extradural,
subdural, and intraparenchymal. Prompt surgical removal of signif-
icant subdural (SDH) and extradural hemorrhage (EDH) is well es-
tablished and widely accepted. Intraparenchymal hemorrhage is more

1
Neurosurgical Trials Group, Newcastle University, Newcastle upon Tyne, United Kingdom.
2
Newcastle Clinical Trials Unit, Newcastle University, Newcastle upon Tyne, United Kingdom.
3
Health Economics Research Unit, University of Aberdeen, Aberdeen, United Kingdom.
4
Department of Medical Physics, South Tees Hospitals Foundation Trust, James Cook University Hospital, Middlesbrough, United Kingdom.
5
Department of Neurosurgery, Heidelberg University, Heidelberg, Germany.
6
Royal Victoria Infirmary, Newcastle upon Tyne, United Kingdom.

A. David Mendelow, Barbara A. Gregson, Elise N. Rowan, Richard Francis, Elaine McColl, Paul McNamee, Iain R. Chambers, Andreas Unterberg,
Dwayne Boyers, and Patrick M. Mitchell 2015; Published by Mary Ann Liebert, Inc. This Open Access article is distributed under the terms of the
Creative Commons License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium,
provided the original work is properly credited.

1312
 SURGERY FOR TRAUMATIC ICH RCT: STITCH(TRAUMA)
common than both these other types and is clearly associated with a
worse outcome, but the role for surgery and its timing remains
undefined. Several terms are used to describe the condition, in-
cluding traumatic intraparenchymal hemorrhage, traumatic intra-
cerebral hemorrhage (TICH), and contusion.
Traditional neurosurgical management of patients with severe
TBI (sTBI) is frequently based on intracranial pressure (ICP) mea-
surement. Patients with high ICP ( > 30 mm Hg) would typically
undergo craniotomy and those with low ICP ( < 20 mm Hg) would be
managed conservatively. Patients with ICP between 20 and 30 mm
Hg would be observed closely and undergo craniotomy if the ICP
rises.5 This ICP-based approach has been recommended by the Brain
Trauma Foundation,6 although publication of the BEST TRIP trial
from Latin America7 has engendered further debate. However, not all
Downloaded from online.liebertpub.com by 36.80.211.182 on 09/24/17. For personal use only.
hospitals have or use ICP monitoring for patients with TICH, even
though they may be classified as having an sTBI. Early management
of patients with TICH requires evaluation to determine whether early
surgery should become part of the standard of care in the same way it
is for significant EDHs8 and SDHs.9
The aim of early surgical TICH removal is to prevent secondary
brain injury, which is thought to be caused by a number of mecha-
nisms. Extravasated blood is believed to be neurotoxic, leading to
secondary injury that may be avoided by early surgical removal.
Journal of Neurotrauma 2015.32:1312-1323.
Larger TICHs may be associated with an ischemic penumbra of brain
tissue that could be salvaged and some TICHs expand to the point
where they cause mass effect, resulting in secondary brain injury.
Contused brain does not seem to recover and appears later as en-
cephalomalacic brain tissue loss on convalescent phase imaging.
Removal of irreversibly damaged brain contusion with TICH does not
increase tissue loss. As with spontaneous intracerebral hemorrhage
(SICH), there are patients who will deteriorate clinically and the
question of early surgery arises to anticipate such secondary damage.
Use of surgery for TICH varies around the world. It is more
frequent in Asia than in Europe or North America. There have been
randomized trials of surgery for SICH,10,11 but none thus far of
surgery for TICH. Patients suffering a TICH tend to be younger
than those suffering a SICH, and therefore level of disability may
have a larger effect on ability to return to employment and eco-
nomic output. TICHs are more likely to be lobar, superficial, and
have a medium-sized volume (2565 cc).12 These differences be-
tween the conditions mean that the role of surgery for TICH cannot
be directly derived from results of the published trials of surgery for
SICH. If early surgery is of benefit to TICH patients, then im-
plementation of early referral and diagnosis with immediate
treatment may reduce incidence of death and disability in this
specific group of TBI patients.
The National Institute for Health and Care Excellence (NICE) in
the UK published their second edition of guidelines for the triage
and management of TBI patients in 2007,13 and the Brain Trauma
Foundation published their guidelines for the surgical management
of TBI in 2006.14 Both of these organizations have emphasized that
studies have been observational and that there is a lack of class 1
evidence from well-designed randomized, controlled trials. Those
unrandomized studies that attempt to compare outcome between
surgical and nonsurgical groups cannot adequately control for
known prognostic variables. The NICE recommended in the 2007
guideline that research is needed to develop a consensus on criteria
for lesions not currently considered to be surgically significant,
namely, TICH. This recommendation facilitated the funding of
[STITCH(TRAUMA)] to find out whether early surgery would
improve outcomes, compared with initial conservative treatment,
in patients with supratentorial TICH.
1313
Methods
Trial design and participants
The STITCH(Trauma) protocol has been published.15 This was an
international, multicenter, prospective, patient-randomized, parallel-
group pragmatic trial comparing early surgical evacuation of TICH
with initial conservative treatment (ISRCTN 19321911, UK NIHR-
HTA grant no.: 07/37/16). Ethical committee favorable opinion was
obtained from the Southampton Multicenter Research Ethics Com-
mittee (REF: 09/H0502/68, June 15, 2009). Local ethical approval was
obtained for each participating center. The trial was conducted ac-
cording to Medical Research Council good clinical practice guidelines.
Formal agreements were in place between the sponsor (Newcastle
upon Tyne NHS Hospitals Foundation Trust), the holder of the study
funding (Newcastle University, Newcastle upon Tyne, UK), and each
participating hospital before commencing the study at each site.
Only TICH patients for whom the treating neurosurgeon was in
equipoise about the benefits of early surgical evacuation, compared
with initial conservative treatment, were eligible for the trial. Pa-
tients considered for the trial had had a computed tomography (CT)
scan to confirm the diagnosis and size as well as location of the
hematoma. Clotting or coagulation problems were corrected before
randomization as per local standard clinical practice. Patients were
included if they were adults within 48 h of TBI and had evidence of
a TICH on CT with a confluent volume of attenuation significantly
raised above that of the background white and gray matter greater
than 10 mL calculated by: (length width height)/2 in cm.16
(Initially, the time criterion was within 24 h of TBI, but subsequent
to an investigators meeting this was increased to allow time for
patients to reach neurosurgery and for the TICH to develop.)
Exclusion criteria were: a significant surface hematoma (EDH or
SDH) requiring surgery; three or more separate hematomas ful-
filling the inclusion criteria; a cerebellar hemorrhage/contusion;
surgery could not be performed within 12 h of randomization; se-
vere pre-existing physical or mental disability or comorbidity that
would lead to a poor outcome even if the patient made a full re-
covery from the TBI; permanent residence outside a study country
preventing follow-up; and if the patient and/or relative expressed a
strong preference for one treatment modality.
Written witnessed informed consent of the patients or their rel-
atives was obtained by neurosurgical staff before randomization.
In total, 59 neurosurgical units in 20 countries completed all
regulatory requirements and registered with the trial.
Interventions
The two trial interventions were early surgery or initial conser-
vative treatment. Early surgery was early evacuation of the he-
matoma by a method of the surgeons choice (within 12 h of
randomization), combined with appropriate best medical treatment.
Initial conservative treatment was best medical treatment combined
with delayed (more than 12 h after randomisation) evacuation if it
became appropriate later. Both groups were monitored according to
local standard neurosurgical practice.
Best medical treatment could include (depending on the prac-
tices within the center) monitoring of ICP or other modalities and
management of metabolism, sodium osmotic pressure, tempera-
ture, and blood gasses.
All patients had a CT scan at 5 days ( 2 days) after ran-
domization to assess changes in hematoma size with and without
surgery.
Information was collected about the status (Glasgow Coma
Score [GCS] and focal signs) of patients through the first 5 days of
their trial progress as well as ICP/CPP (cerebral perfusion pressure)
measures (in invasively monitored patients), to describe any change
in status that led to a change in equipoise for the treating neuro-
surgeon, and subsequent surgery in patients initially randomized
to conservative treatment. At 2 weeks after randomization or at
 1314
discharge or death (whichever occurred first), a discharge/2-week
form was completed by the responsible neurosurgeon or research
nurse. This form recorded the patients status at that time, the
mechanism of injury, whether and when they had surgery (in-
cluding why, if randomized to initial conservative treatment, or
why not, if randomized to early surgery), the GCS and localizing
features for the 5 days after randomisation, the occurrence of any
adverse events (AEs) after randomization (including death, pul-
monary embolism, deep vein thrombosis, and surgical site infec-
tion), and past medical history.
Before assessing outcome each patients GP (in the UK) or
consultant (outside the UK) was contacted to check that the patient
was alive, confirm the patients place of residence, and complete a
major AEs form.
Downloaded from online.liebertpub.com by 36.80.211.182 on 09/24/17. For personal use only.
Randomization and masking
To minimise biases resulting from lack of concealment of allo-
cation, randomization was undertaken centrally by an independent
24-h telephone and Web randomization service based in Aberdeen
University (Aberdeen, UK). Allocation was stratified by geograph-
ical region, with a minimization algorithm based on age group, and
severity (as measured by whether or not the pupils were equal and
reacting), with a random component (i.e., with probability of 80%).
Throughout the study, data broken down by treatment assign-
ment were never provided to individual investigators or to the study
Journal of Neurotrauma 2015.32:1312-1323.
team. Outcome was assessed by questionnaires to patients. If it was
necessary to administer these questionnaires in person, then the
interviewer was blind to treatment allocation. Outcome assign-
ment, data cleaning, and CT assessment were all made before un-
blinding the treatment assignment.
Outcomes
Outcome was assessed at 6 months by a postal questionnaire
translated into the appropriate languages and mailed to patients or
their relatives/carers. In those centers where the postal systems
were problematic, or there were literacy or language problems,
questionnaires could be completed by a social worker or research
nurse, who did not know the treatment allocation, in interview with
the patient or relative.
Primary outcome was based on the 6-month Glasgow Outcome
Scale (GOS) dichotomized into favorable and unfavorable outcome;
dead, vegetative, and severe disability were coded as unfavorable and
moderate disability and good recovery as favorable. Secondary
outcomes were mortality, time to death, extended Glasgow Outcome
Scale (GOSE), Rankin, and European Quality of Life Five Dimen-
sion Scale (EQ-5D) at 6 months. Crossover and major event rates
(death, pulmonary embolism or deep vein thrombosis, infection, and
rehemorrhage) in each treatment group were also reported.
Statistical analysis
Previous observational studies have suggested a favorable out-
come in the nonoperated group of approximately 40% and a fa-
vorable outcome in the surgical group of approximately 6070%.
Target sample size was 840 (420 in each arm). This was calculated
assuming a more conservative favorable outcome (good recovery
or moderate disability on the GOS) of 50% from conservative
treatment, 10% benefit (i.e., 50% vs. 60%) from surgery, 5% sig-
nificance with 80% power, and a safety margin built in to allow for
loss to follow-up, making the total target sample size 840 patients
(420 randomized to each treatment arm).
The independent data monitoring committee (DMC) reviewed
data from the study after 50, 100, and 150 patients had been recruited.
These interim reviews were confidential to only the data manager and
the DMC. The trial was only to be stopped early by the DMC if one or
other treatment policy showed an advantage at a very high signifi-
cance level, or if recruitment rates were unexpectedly low. The DMC
MENDELOW ET AL.
recommended, at each review, that the trial should continue. How-
ever, in February 2012, the funding agency decided to halt this in-
ternational study with effect from the end of September 2012 for
failure to recruit in the UK. At this time, 6 patients had been
recruited from the UK, despite our actively encouraging UK center
participation throughout the trial. In total, 10 of the possible 36 UK
centers expressed written interest in the study, but only seven com-
pleted all regulatory requirements to join the trial. Considerable ef-
fort was expended in trying continuously to recruit UK centers and
raise the profile of the studythe study team had a high profile at the
national neurosurgical meetings and national neurosurgical research
meetings giving presentations about progress on the study and
manning a stand where investigators could discuss the study with the
team. Monthly e-mail newsletters were sent to neurosurgical de-
partments and paper copies were posted quarterly. Talks were given
in individual centers to encourage recruitment. Help with all regu-
latory submissions was also provided. Despite these considerable
efforts, only 6 patients could be recruited in the UK.
Analysis was on an intention-to-treat (ITT) basis. The analysis
plan was adapted, following the decision from the funding agency to
halt the trial early, agreed by the trial steering committee and pub-
lished on the study website before unblinding the data. The primary
analysis was a simple categorical frequency comparison using the
uncorrected chi-squared test for favorable and unfavorable outcomes
at 6 months.17,18 A sensitivity analysis using logistic regression was
undertaken to adjust for age, volume of hematoma, and GCS.
Secondary outcome analyses included proportional odds model
analysis of GOS, GOSE, and Rankin at 6 months, Kaplan-Meiers
survival curve with log-rank test, and mortality. For dichotomized
outcomes, absolute differences and 95% confidence intervals (CIs)
were reported.
Minimal subgroup analysis was undertaken and regarded as
exploratory. Odds ratios (ORs) and 95% CIs were reported for the
following subgroups: age (two bands using randomization strata:
< 50, 50, or moregiven that there were very few patients over 70
years of age, the two upper age bands were combined); volume of
hematoma (using median split < = 23 and > 23 mL); GCS (using
standard classification of TBI severe, moderate, or minor: 38,
912, or 1315); time from ictus to randomization (using median
split < 21 and > = 21 h); and geographical region (four bands:
Europe, India, China, and other). Interaction tests (chi-squared test
for heterogeneity) were undertaken and relevant p values reported.
Costing analysis
The costing analysis was undertaken on the ITT basis from an
international health services perspective. Resource-use require-
ments to deliver the interventions (e.g., staff time and overheads)
and time spent on hospital wards were collected using site-specific
questionnaires and case report forms. Hospital readmissions were
reported on participant outcome questionnaires. Costing followed
recommended procedures for international studies,19,20 applying
country-specific unit costs (sourced from site-specific question-
naires) to resource-use data to generate total costs. Costs were
transformed into 2013 international dollars21 and reported as mean
(standard deviation; SD) for each treatment group.
A generalized linear regression model (GLM), specifying a
gamma family and identity link, was used to estimate the impact of
treatment allocation on costs accounting for skewed data and ad-
justing for patient characteristics (age and sex). Sensitivity analyses
explored the use of alternative models for analyzing costs. Sub-
group analysis presented raw mean costs according to World Bank
classifications,22 based on gross national income (GNI) as follows:
low income (GNI < = $1,005, e.g., Nepal); lower middle income
(GNI $1,006$3,975, e.g., India); upper middle income (GNI
$3,976$12,275, e.g., China); and high income (GNI > = f
$12,276, e.g., western Europe and United States). Owing to small
sample sizes, regression analyses were not undertaken.
 SURGERY FOR TRAUMATIC ICH RCT: STITCH(TRAUMA) 1315

Role of the funding source Results

Neither the sponsor nor the funder of the study had any role
in study design, data gathering, analysis and interpretation,
or writing the report. The corresponding author, E.N.R., and
R.F. had full access to all the data in the study, and all members
of the writing committee had responsibility for the decision to
submit.
Downloaded from online.liebertpub.com by 36.80.211.182 on 09/24/17. For personal use only.
Journal of Neurotrauma 2015.32:1312-1323.
One hundred and seventy patients were recruited from 31 centers
in 13 countries between December 2009 and September 2012 and
randomly assigned to treatment groups: 83 to early surgery and 87
to initial conservative treatment. Two patients were excluded be-
cause the treatment decision was made before randomization: In
one case, the patient had surgery before randomization and in the

FIG. 1. Flow chart for STITCH(Trauma) patients. *One site recruited 1 patient, but had undertaken surgery before randomization
the patient was allocated to initial conservative treatment; another site recruited 1 patient, for whom a treatment decision not to operate
was made before the patient was randomizedthis patient was allocated to early surgery. Because of the severe breach of protocol,
these patients were excluded.
 1316 MENDELOW ET AL.

other an early decision was made not to operate (Fig. 1). These were
serious protocol violations. All other patients were included in the
analysis, which therefore reports results for 82 patients assigned to
early surgery and 86 assigned to initial conservative treatment.
Table 1 shows the distribution of baseline variables between the
two treatment groups. Patients ranged in age from 16 to 83, with a
median age of 50, and 122 (73%) were male. Before the TBI, 164
(98%) were Rankin 0 or 1, and 22 (13%) had a history of cardio-
vascular disease. The main causes of the TBI were road traffic acci-
dents (RTAs; 113; 67%) and falls (47; 28%). Most of the RTAs were
motorbike riders (45; 40%) or pedestrians (29; 26%). Sixty-eight
patients (40%) were admitted to another hospital before their transfer
to the neurosurgical unit. At the time of randomization, 70 (42%)
patients had a GCS of 1315, 71 (42%) a GCS of 912, and 27 (16%) a
Downloaded from online.liebertpub.com by 36.80.211.182 on 09/24/17. For personal use only.
GCS of 8 or less. The volume of the largest hematoma varied between
10 and 97 mL, with a median of 23 mL, and 61 (36%) patients had a
second hematoma between 0 and 26 mL, with median of 3 mL. There
were no differences in the three individual components of the GCS,
handedness, or characteristics of the second hematoma.
Of the 82 patients in the early surgery group, only 61 (74%) had
surgery, 57 (93%) of these within 12 h of randomisation (see Table
2). The reasons for not having surgery were patient or relative
refusal (15), improvement (1), deterioration (2), seizures (1), an-
esthetic risk (1), and change of history suggesting SICH rather than
TICH (1). Although informed consent was obtained before ran-
domization, patients often had more than one relative and further
discussion could lead to a change of opinion. Of the 15 patients/
relatives who refused surgery, 9 were in China and 5 in India. Of the

86 patients randomized to initial conservative treatment, 31 (36%)

Table 1. Baseline Variables

Table 2. Surgery Details for Early Surgery Patients
Early surgery Initial conservative Who Had Surgery and Initial Conservative Patients
Variable (N = 82) treatment (N = 86) Who Required Delayed Surgery
Age (years), Early Initial
median (IQR) range 51 (3263) 1883 50 (3361) 1677 surgery conservative
Mean (SD) 48 (17.7) 48 (16.9) surgical treatment
Journal of Neurotrauma 2015.32:1312-1323.

Age band (%) cases surgical cases

< 50 37 (45) 42 (49) (N = 61; 74%) (N = 31; 36%)
5069 34 (42) 33 (38)
70 + 11 (13) 11 (13) Method (%)
Craniotomy 59 (97) 25 (81)
Sex (%) Other 2 (3) 6 (19)
Male 57 (70) 65 (76)
Female 25 (30) 21 (24) Bone flap replaced (%) 47 (77) 13 (42)
Other cranial surgery (%) 1 (2) 3 (10)
GCS total (%) Paralyzed and sedated (%) 17 (28) 12 (39)
3 0 (0) 1 (1) Any noncranial surgery (%) 1 (2) 2 (7)
4 0 (0) 0 (0)
5 1 (1) 2 (2) Preoperative GCSeye (%)
6 6 (7) 3 (3) 1 5 (8) 15 (48)
7 4 (5) 3 (3) 2 18 (30) 8 (26)
8 1 (1) 6 (7) 3 19 (31) 5 (16)
9 11 (13) 8 (9) 4 19 (31) 3 (10)
10 11 (13) 14 (16) Preoperative GCSVerbal (%)
11 6 (7) 8 (9) 1 13 (21) 16 (52)
12 6 (7) 7 (8) 2 15 (25) 7 (23)
13 10 (12) 8 (9) 3 6 (10) 5 (16)
14 14 (17) 13 (15) 4 18 (30) 0 (0)
15 12 (15) 13 (15) 5 9 (15) 3 (10)
Pupils (%) Preoperative GCSMotor (%)
Both reactive 77 (94) 79 (92) 1 0 (0) 4 (13)
One reactive 3 (4) 3 (3) 2 2 (3) 1 (3)
Both unreactive 2 (2) 4 (5) 3 6 (10) 3 (10)
Volume of largest 25 (1837) 1196 23 (1532) 1097 4 4 (7) 6 (19)
hematoma (mL) Mean = 31 (18.0) Mean = 27 (16.8) 5 26 (43) 14 (45)
6 23 (38) 3 (10)
Location of largest hemorrhage (%)
Frontal 36 (44) 43 (50) Time randomisation 3 (16) < 124 25 (679) < 1318
Temporal 39 (48) 37 (43) to surgery (h) Mean: 4 (4.5) Mean: 58 (75.6)
Parietal 4 (5) 5 (6) Surgery within 12 h 57 (93) 10 (32)
Occipital 3 (4) 1 (1) of randomization (%)
Second hematoma 28 (34) 33 (38) Time injury to 23 (1636) 469 45 (2699) 9332
present (%) surgery (h) Mean 26 (13.8) Mean 78 (79.0)
Time to 21 (1331) 348 22 (1428) 448 Surgery within 9 (15) 3 (10)
randomization (h) Mean = 22 (11.7) Mean = 22 (10.6) 12 hours of injury (%)

For continuous variables, median (quartiles) and range are presented For continuous variables, median (quartiles) and range are presented
plus mean and SD; for categorical variables, the number of cases plus mean and standard deviation; for categorical variables, number of
(percentage) is presented. cases (percentage) are presented.
IQR, interquartile range; SD, standard deviation; GCS, Glasgow Coma Score. GCS, Glasgow Coma Score.
 SURGERY FOR TRAUMATIC ICH RCT: STITCH(TRAUMA)
had surgery within 14 days of randomization, 10 (32%) of these
within 12 h. The reasons for having surgery were neurological
deterioration (29), no shrinkage in hematoma size (1), and rise in
ICP (1). Neurological deterioration was identified by a drop in
GCS, enlargement of the hematoma or increase in mid-line shift,
increase in weakness, or change in pupil size or reactivity.
Surgical patients in the early surgery group were more likely to
have craniotomy than those in the initial conservative group (97%
vs. 81%; p = 0.016, Fishers test). One patient in the initial con-
servative group had burrhole surgery, but all others who did not
have craniotomy had craniectomy. The bone flap was more likely
to be replaced in the surgical patients in the early surgery group
(77%) than in the initial conservative group (42%; Fisher tests,
p = 0.001). As Table 2 demonstrates, surgical patients in the early
Downloaded from online.liebertpub.com by 36.80.211.182 on 09/24/17. For personal use only.
surgery group had significantly higher preoperative GCS scores
for all subscales than those requiring surgery in the initial con-
servative group. Comparison of the baseline characteristics of
patients in the initial conservative group who had surgery, with
those that did not, showed that patients who deteriorated and went
on to have surgery had larger hematomas initially (Mann-Whit-
neys test, p = 0.010) and were more likely to have at least one
pupil unreactive (Fishers test, p = 0.0005), but did not differ on
age, GCS at the time of randomization, or presence of a second
Journal of Neurotrauma 2015.32:1312-1323.
hematoma.
At 2 weeks postrandomization, there were similar proportions of
patients in the two groups who were still on the neurosurgical ward:
29 (35%) of the early surgery patients and 32 (37%) of the initial
conservative patients. Further, similar proportions had been trans-
ferred to another ward or hospital (3 [4%] and 4 [5%], respectively).
However, 43 early surgery patients (52%) had been discharged,
compared to 33 of the initial conservative patients (38%). Further,
there was a significant difference in the percentage that had died by
2 weeks: 7 (9%) early surgery patients compared to 17 (20%) initial
conservative patients (Fishers test, p = 0.047). At some point in the
first 2 weeks, 7 (9%) early surgery patients were ICP monitored
compared to 16 (19%) initial conservative patients ( p = 0.073), and
this affected management decisions in 1 early surgery patient
compared to 10 initial conservative patients ( p = 0.069). Patients
were less likely to be monitored in India, where the ICP monitoring
rate was 4% (3 of 74), compared to 21% elsewhere (20 of 94). Very
few postrandomization events were recorded during the first 2
weeks of the hospital stay: Pneumonia was reported in 8 early
surgery patients and 8 initial conservative patients, ischemic stroke
(0 and 1), pulmonary embolism (1 and 2), postoperative extradural
(0 and 2), septicemia (1 and 0), urinary tract infection (1 and 0),
seizures (3 and 0), and other (5 and 1).
Primary outcome
Six-month outcome was available for 82 early surgery patients
and 85 initial conservative patients; 1 patient from the initial con-
servative group was lost to follow-up. Fifty-two (63%) early sur-
gery patients had a favorable outcome on the dichotomized GOS,
compared to 45 (53%) initial conservative patients (OR, 0.65; 95%
CI, 0.35, 1.21; p = 0.171); an absolute difference of 10.5% (95%
CI, - 4.4, 25.3; see table 3; Fig. 2). Adjusting for age, volume, and
GCS gives an OR of 0.58 (95% CI, 0.29, 1.16; p = 0.122).
Secondary outcomes
However, there was a highly significant difference in mortality at
6 months, with 12 (15%) early surgery patients dying compared to
28 (33%) initial conservative patients (OR, 0.35; 95% CI, 0.16,
1317
0.75; p = 0.007), for an absolute difference of 18.3% (95% CI, 5.7,
30.9). Figure 3 shows the Kaplan-Meiers plot of survival for the
two groups of patients, illustrating the significant survival advan-
tage of early surgery compared with initial conservative treatment
( p = 0.008). Table 3 and Figure 2 show the distribution of GOS,
GOSE, and Rankin at 6 months by treatment group. For each of
these secondary outcomes, there is a significant trend in better
outcome in the early surgery group (chi-squared trend: p = 0.047,
p = 0.052, and p = 0.043 respectively), although the proportional
odds models did not reach statistical significance (OR, 0.67; 95%
CI, 0.39, 1.16; p = 0.153; OR, 0.66, 95% CI, 0.38, 1.13; p = 0.127;
OR, 0.67; 95% CI, 0.39, 1.15; p = 0.147).
The main causes of death were the initial head injury (5 early
surgery, 14 initial conservative) and pneumonia (4 early surgery, 2
initial conservative). Other causes in the initial conservative
treatment group included cachexia (2), ischemic stroke (2), men-
ingitis (1), pulmonary embolism (2), renal (1), TBI and surgery (1),
seizure (1), and unknown sudden death in the community (1). In
the early surgery group, the other causes were hypovolemic shock
(1), pulmonary embolism (1), TBI and surgery (1), and unknown in
the community (1). Only 8 non-death-related major AEs were re-
corded: seizure (3); new/enlarged hematoma (2); infection (2); and
other (1).
Prespecified subgroup analyses are shown in Figure 4. None of
the subgroups displayed any significant heterogeneity of treatment
response, although the patients with a GCS of 912 had the best
response from early surgery.
Among the patients who were allocated to early surgery and
had surgery, 33% (20 of 61) died or were severely disabled at 6
months. However, 65% (20 of 31) of patients who were allocated
to initial conservative treatment and had delayed surgery died or
were severely disabled at 6 months, whereas 37% (20 of 54) of the
conservative patients who did not have surgery had an unfavorable
outcome.
An unadjusted comparison of raw mean costs showed that early
surgery was, on average, $476 more costly than conservative
management (Table 4). GLM regression analysis, adjusting for
patient characteristics, showed early surgery to be $1,774 more
costly (95% CI, - $284, $3,831) than initial conservative treat-
ment. Sensitivity analyses showed that overall conclusions were
robust to the choice of regression model for the analysis. Results
from subgroup analyses were highly uncertain based on small
sample sizes (and too small to conduct regression analysis) and
should therefore be interpreted with caution.
Discussion
Although the trial was stopped early by the UK NIHR-HTA with
an associated reduction in statistical power, there were some clin-
ically significant results. These included a statistically significant
survival advantage (85% vs. 67%) and a nonsignificant benefit on
GOS, both associated with early surgery.
Early management of patients with TICH is not harmonized
around the world. Timing of surgery in patients with parenchymal
hematomas post-TBI has not been standardized. This contrasts with
patients who develop EDHs or acute SDHs, because guidelines
(NICE 2nd edition) based on strong observational data8,9 have re-
commended early and expeditious scanning and surgery. Not all
TICHs require removal and neither do all the contusions associ-
ated with them. Generally, clinical deterioration and expansion of
the hematomas and their associated edema tend to trigger the need
for surgery. If it were possible to anticipate these changes, then
 1318 MENDELOW ET AL.

Table 3. Outcomes at 6 Months

Test and p value

Early surgery Initial conservative treatment Absolute difference (95% CI)

Primary outcome (%) N = 82 N = 85

Unfavorable 30 (37) 40 (47) v2 p = 0.170
Favorable 52 (63) 45 (53) 10.5 ( - 4.425.3)
Secondary outcomes N = 82 N = 85
Mortality at 6 months (%)
Dead 12 (15) 28 (33) v2 p = 0.006
Alive 70 (85) 57 (67) 18.3 (5.730.9)
Rankin (%)
Unfavorable 27 (33) 37 (44) v2 p = 0.159
Favorable 55 (67) 48 (56) 10.6 ( - 4.025.3)
Downloaded from online.liebertpub.com by 36.80.211.182 on 09/24/17. For personal use only.

GOS (%)
Dead 12 (15) 28 (33) v2 trend p = 0.047
Vegetative 0 (0) 0 (0)
Severely dependent 18 (22) 12 (14) POM p = 0.153
Moderately dependent 26 (32) 18 (21)
Good recovery 26 (32) 27(32)
GOSE (%)
Dead 12 (15) 28 (33) v2 trend p = 0.052
Vegetative 0 (0) 0 (0)
Lower SD 4 (5) 8 (9) POM p = 0.127
Journal of Neurotrauma 2015.32:1312-1323.

Upper SD 14 (17) 4 (5)

Lower MD 5 (6) 3 (4)
Upper MD 21 (26) 15 (18)
Lower GR 12 (15) 12 (14)
Upper GR 14 (17) 15 (18)
Rankin (%)
0 17 (21) 18 (21) v2 trend p = 0.043
1 27 (33) 22 (26)
2 11 (13) 8 (9) POM p = 0.147
3 8 (10) 4 (5)
4 7 (9) 3 (4)
5 0 (0) 2 (2)
Dead 12 (15) 28 (33)
EuroQoL Index
Median 0.80 0.71 M-W p = 0.218
Quartiles 0.521.00 0.001.00
Range - 0.331.00 - 0.591.00
Limb movement (%)
Worst affected lega
Unaffected 50 (72) 47 (82) v2 0.374
Weak 18 (26) 9 (16)
Paralysed 1 (1) 1 (2)
Worst affected arma
Unaffected 48 (70) 43 (75) v2 0.464
Weak 21 (30) 14 (25)
Paralysed 0 (0) 0 (0)

Number of cases (percentage) are presented; EuroQol utility index calculated using UK weightings provided by the EuroQol Group Foundation; tests
conducted were v2 (chi-squared), v2 trend (chi-squared for trend), POM (proportional odds model), and M-W (Mann-Whitney). For each test, the p value
is given. Absolute differences with 95% confidence intervals are presented for binary outcomes.
a
One patient did not provide information about their limb movements.
GOS, Glasgow Outcome Scale; GOSE, Glasgow Outcome Scale Extended; SD, severe disability; MD, moderate disability; GR, good recovery.

secondary brain damage would be avoided. The objective of the
STITCH(TRAUMA) trial was to discover whether early surgery
would prevent the secondary deterioration so often observed with
conservative treatment. Though the primary outcome is not sta-
tistically significant, there is a strong signal that early surgery will
indeed prevent such deterioration and save lives. This is noted in
the highly significant reduction in mortality and the better outcomes
in the ordinal analysis of the GOS and Rankin scales. A larger trial
is urgently needed to confirm or refute this signal, which is par-
ticularly strong in patients with a randomization GCS of 912. This
same group of patients with SICH had the best outcomes with early
surgery in a previous study.23
Some units in some countries routinely measure ICP whereas
others do not. In this study, 86% of patients were not monitored for
ICP either because the hospital did not have the technology avail-
able or because they do not routinely use it for this patient group.24
The question of early surgery for TICH may be particularly im-
portant in those countries that do not measure ICP. An analysis of
global ICP utilization trends and aspirations has confirmed that
many countries in the world do not have the facilities and resources
 SURGERY FOR TRAUMATIC ICH RCT: STITCH(TRAUMA)
Downloaded from online.liebertpub.com by 36.80.211.182 on 09/24/17. For personal use only.
Journal of Neurotrauma 2015.32:1312-1323.
FIG. 2. Outcome at 6 months. Statistical significance tests for
outcome. (A) Proportional odds model, p = 0.153; chi-squared for
trend, p = 0.047; outcome. (B) Proportional odds model, p = 0.127;
chi-squared for trend, p = 0.052; outcome. (C) Proportional odds
model, p = 0.147; chi-squared for trend, p = 0.043. ES, early sur-
gery; ICT, initial conservative treatment.
for ICP monitoring.25 It is in these very countries that this STITCH
trial is most relevant to the question: Should early surgery be un-
dertaken in patients with TICH when there is no option for ICP
measurement?
There were crossovers from initial conservative treatment to early
surgery and vice versa, as occurs in all surgical trials. This is because
surgeons feel compelled to provide rescue surgery to those patients
1319
FIG. 3. Kaplan-Meiers survival analysis. Log-rank test,
p = 0.0081.
randomized to initial conservative treatment who subsequently de-
teriorate. On the other hand, some patients who were randomized to
early surgery did not have surgery because their families withdrew
consent. Despite these crossovers, the absolute benefit of early sur-
gery exceeded 10% and was almost statistically significant. If the
total of 840 planned patients had been recruited and if the same trend
had transpired, this would have been a statistically significant result.
In addition, the patients who had delayed surgery had deteriorated to
a much poorer clinical state and this was associated with a much
poorer outcome (65% dead or severely disabled, compared to only
33% in those operated upon early). This observation supports the
primary hypothesis that early surgery is advantageous for TICH.
An analysis of outcome by whether patients actually had surgery
or not is complex and biased because the decision for surgery was
guided by the study in some cases and by a change in status in
others: Some patients would have surgery before deterioration and
others would only have surgery after deterioration. In the non-
surgery group, some patients did not have early surgery mainly
because the relative refused, whereas others did not have surgery
because that was the allocation (and either they did not deteriorate
or if they did their surgeon or anesthetist did not decide to take them
to surgery). We will publish a separate article that will address the
driving forces behind the crossovers. We will also publish a sepa-
rate description of the CT characteristics at baseline and at 5 days in
those patients who survived.
Predicting which patients will deteriorate is complex, and the
Surgical Trial in Intracerebral Haemorrhage (STICH II) identified a
small number of patients (GCS between 9 and 12) that may benefit
from such anticipatory treatment.23 In general, SICH patients with
a good prognosis (GCS between 13 and 15) can be safely observed
and only require craniotomy if they deteriorate. This is because there
is enough time to perform a craniotomy before other secondary
mechanisms, such as brain edema, mass effect with herniation, and
reduced CPP from elevated ICP, cause harm. This may also be true
for TICH patients. In particular, those TICH patients with an initial
 1320
Downloaded from online.liebertpub.com by 36.80.211.182 on 09/24/17. For personal use only.
Journal of Neurotrauma 2015.32:1312-1323.
FIG. 4. Subgroup analysis. M-H, Mann-Whitney; CI, confidence interval; GCS, Glasgow Coma Score.
GCS between 9 and 12 had the best outcome with early surgery. The
economic analysis indicates that a strategy of early surgery is asso-
ciated with a nonsignificant increase in health care costs. Further
work will estimate the cost-effectiveness of early surgery versus
conservative management using 6-month and 1-year follow-up data.
This will demonstrate the magnitude of additional costs over the
longer term, as well as whether any additional costs are associated
with sufficient benefits in terms of improvements in quality of life,
measured by the EQ-5D, and length of life.
MENDELOW ET AL.
Conclusion
A larger trial is needed to confirm this potentially very beneficial
effect of earlier surgery. In the interim, there is a strong case
for operating on patients with TICH who have a GCS of 912.
Those who are alert or just confused (GCS 1315) can probably
be watched carefully for any deterioration because there is a
safety margin, which diminishes the lower down the GCS the pa-
tient descends. Once the GCS has descended below 9, surgical
 SURGERY FOR TRAUMATIC ICH RCT: STITCH(TRAUMA) 1321

Table 4. Costing Analysis (All Countries) and by Country Income Group

Early surgery (N = 82) Initial conservative treatment (N = 86)

Resource use; Costs; Resource use; Costs;

All countries mean (SD) mean (SD) mean (SD) mean (SD) Difference of means

Cost surgery 981 (1678) 515 (1206)

Cost ICU 4.18 (4.2) 2808 (5762) 4.06 (4.61) 2988 (6131)
Cost HDU 1.72 (2.55) 385 (1053) 1.76 (3.01) 461 (1445)
Cost ward 11.88 (15.95) 3595 (10,206) 14.24 (29.43) 3997 (13,789)
Cost readmission 4.23 (14.43) 1145 (5775) 2.42 (9.63) 421 (1720) + 476 GLM model + 1774
Total cost 8812 (18,032)a 8336 (18,685)a (95% CI, - 2843831)

Low-income countries Early surgery (N = 6) Initial conservative treatment (N = 10) Difference of means
Downloaded from online.liebertpub.com by 36.80.211.182 on 09/24/17. For personal use only.

Cost surgery 142 (0) 14 (45)

Cost ICU 0.83 (1.60) 203 (391) 1.2 (2.7) 293 (659)
Cost HDU 3.83 (0.75) 468 (92) 3.5 (2.12) 427 (259)
Cost ward 5.33 (1.03) 325 (63) 6.3 (6.43) 384 (392)
Cost readmission 0 (0) 0 (0) 0 (0) 0 (0)
Total cost 1139 (418) 1118 (614) + 20

Low-middle-income Early surgery (N = 40) Initial conservative treatment (N = 39) Difference of means
countries
Journal of Neurotrauma 2015.32:1312-1323.

Cost surgery 439 (511) 176 (369)

Cost ICU 3.2 (3.78) 580 (1449) 2.38 (3.39) 227 (314)
Cost HDU 1.93 (2.84) 87 (128) 1.97 (3.08) 168 (513)
Cost ward 5.7 (5.84) 64 (98) 6.31 (5.29) 125 (364)
Cost readmission 0.35 (2.21) 3 (20) 0.13 (0.59) 1 (5)
Total cost 1174 (1583) 697 (964) + 477

Upper-middle-income Early surgery (N = 28) Initial conservative treatment (N = 30) Difference of means
countries

Cost surgery 1089 (1174) 822 (1031)

Cost ICU 5.43 (3.79) 4272 (4134) 6.93 (4.49) 6010 (5588)
Cost HDU 0.93 (1.84) 643 (1261) 1.17 (3.26) 821 (2295)
Cost ward 16.86 (16.30) 3603 (4132) 14.27 (26.05) 3080 (6267)
Cost readmission 8.39 (20.55) 997 (2986) 6.23 (15.48) 805 (1881)
Total cost 10,603 (7517) 11,538 (10,149) - 936

High-income countries Early surgery (N = 8) Initial conservative treatment (N = 7) Difference of means

Cost surgery 4927 (3617) 2020 (3542)

Cost ICU 7.25 (5.95) 13,432 (14,847) 5.14 (6.72) 10,310 (15,989)
Cost HDU 1.88 (3.18) 1089 (2668) 0.57 (0.98) 622 (964)
Cost ward 30.25 (31.45) 23,671 (24,462) 69.71 (68.18) 34,662 (35,806)
Cost readmission 12.27 (22.53) 8233 (16,895) 2.29 (6.05) 1719 (4547)
Total cost 46,489 (38,880)a 47,483 (46,221)a - 994
a
Total mean cost is not equal to the sum of the resource use. This is because of the use of DRG costs per episode of care, applied to resource use in
Germany.
SD, standard deviation; ICU, intensive care unti; HDU, high dependency unit; GLM, generalized linear regression model; CI, confidence interval.

intervention appears to be less effective. A strategy of early sur-
gery is associated with a small, nonsignificant increase in health
care costs, but further analysis using longer-term follow-up
data are required to establish better estimates of costs and cost-
effectiveness.
Acknowledgments
This project was funded by the National Institute of Health
Research (NIHR) Health Technology Assessment program (project
no.: 07/37/16). The views and opinions expressed therein are those
of the authors and do not necessarily reflect those of the HTA
program, NIHR, NHS, or the Department of Health.
Author Disclosure Statement
A.D.M. is director of the Newcastle Neurosurgery Foundation,
which is a nonprofit organization for academic research and edu-
cation, and he is an adviser to Stryker (craniofacial surgery com-
mittee). B.A.G., E.N.R., and R.F. received salary support from the
STITCH(Trauma) grant, but not from any commercial organiza-
tions. All surgeons in fee-for-service health care systems receive
additional fees for undertaking surgery.
 1322
References
1. National Institute for Health and Care Excellence. (2003). NICE
clinical guideline sets out recommendations for NHS care of people
who have suffered a head injury. National Institute for Health and
Care Excellence: London, UK.
2. Fearnside, M., and Simpson, D. (1997). Epidemiology, in: Head In-
jury. P. Reilly (ed). Chapman & Hall: London, pps. 123.
3. Tagliaferri, F., Compagnone, C., Korsic, M., Servadei, F., and Kraus,
J. (2006). A systematic review of brain injury epidemiology in Europe.
Acta Neurochir. (Wien) 148, 255268; discussion, 268.
4. Gabbe, B.J., Biostat, G.D., Lecky, F.E., Bouamra, O., Woodford, M.,
Jenks, T., Coats, T.J., and Cameron, P.A. (2011). The effect of an
organized trauma system on mortality in major trauma involving se-
rious head injury: a comparison of the United kingdom and victoria,
australia. Ann. Surg. 253, 138143.
5. Gallbraith, S., and Teasdale, G. (1981). Predicting the need for op-
Downloaded from online.liebertpub.com by 36.80.211.182 on 09/24/17. For personal use only.
eration in the patient with an occult traumatic intracranial hematoma.
J. Neurosurg. 55, 7581.
6. Brain Trauma Foundation; American Association of Neurological
Surgeons; Congress of Neurological Surgeons; Joint Section on
Neurotrauma and Critical Care, AANS/CNS, Bratton, S.L., Chestnut,
R.M., Ghajar, J., McConnell Hammond, F.F., Harris, O.A., Hartl, R.,
Manley, G.T., Nemecek, A., Newell, D.W., Rosenthal, G., Schouten,
J., Shutter, L., Timmons, S.D., Ullman, J.S., Videtta, W., Wilberger,
J.E. and Wright, D.W. (2007). Guidelines for the management of
severe traumatic brain injury. VIII. Intracranial pressure thresholds. J.
Neurotrauma 24, Suppl. 1, S55S58. Erratum in: J. Neurotrauma
Journal of Neurotrauma 2015.32:1312-1323.
(2008) 25, 276278.
7. Chesnut, R.M., Temkin, N., Carney, N., Dikmen, S., Rondina, C.,
Videtta, W., Petroni, G., Lujan, S., Pridgeon, J., Barber, J., Machamer,
J., Chaddock, K., Celix, J.M., Cherner, M., and Hendrix, T. (2012). A
trial of intracranial-pressure monitoring in traumatic brain injury. N.
Engl. J. Med. 367, 24712481.
8. Mendelow, A.D., Karmi, M.Z., Paul, K.S., Fuller, G.A., and Gilling-
ham, F.J. (1979). Extradural haematoma: effect of delayed treatment.
Br. Med. J. 1, 12401242.
9. Seelig, J.M., Becker, D.P., Miller, J.D., Greenberg, R.P., Ward, J.D.,
and Choi, S.C. (1981). Traumatic acute subdural hematoma: major
mortality reduction in comatose patients treated within four hours. N.
Engl. J. Med. 304, 15111518.
10. Prasad, K., Mendelow, A.D., and Gregson, B. (2008). Surgery for
primary supratentorial intracerebral haemorrhage. Cochrane Database
Syst. Rev. (4), CD000200.
11. Mendelow, A., and Gregson, B. (2011). Surgery for intracerebral
hemorrhage, in: Stroke: Pathophysiology, Diagnosis and Manage-
ment. J. Mohr, P. Wolf, and J.C. Grotta (eds). WB Saunders: Phila-
delphia, PA.
12. Siddique, M.S., Gregson, B.A., Fernandes, H.M., Barnes, J., Tread-
well, L., Wooldridge, T.D., and Mendelow, A.D. (2002). Comparative
study of traumatic and spontaneous intracerebral hemorrhage. J.
Neurosurg. 96, 8689.
13. National Institute of Health and Care Excellence. (2007). Clinical
guidelines CG56. Head Injury: triage, assessment, investigation and
early management of head injury in infants, children and adults. Na-
tional Institute of Health and Care Excellence: London, UK.
14. Bullock, M., Chesnut, R., Ghajar, J., Gordon, D., Hartl, R., Newell, D.,
Servadei, F., Walters, B., and Wilberger, J. (2006). Surgical man-
MENDELOW ET AL.
agement of traumatic parenchymal lesions. Neurosurgery 58, S2546;
discussion, Siiv.
15. Gregson, B.A., Rowan, E.N., Mitchell, P.M., Unterberg, A., McColl,
E.M., Chambers, I.R., McNamee, P., and Mendelow, A.D. (2012).
Surgical trial in traumatic intracerebral hemorrhage (STITCH(-
Trauma)): study protocol for a randomized controlled trial. Trials
13, 193.
16. Broderick, J.P., Brott, T.G., Duldner, J.E., Tomsick, T., and Huster, G.
(1993). Volume of intracerebral hemorrhage. A powerful and easy-to-
use predictor of 30-day mortality. Stroke 24, 987993.
17. Murray, G.D., Barer, D., Choi, S., Fernandes, H., Gregson, B., Lees,
K.R., Maas, A.I., Marmarou, A., Mendelow, A.D., Steyerberg, E.W.,
Taylor, G.S., Teasdale, G.M., and Weir, C.J. (2005). Design and
analysis of phase III trials with ordered outcome scales: the concept of
the sliding dichotomy. J. Neurotrauma 22, 511517.
18. Mendelow, A.D., Gregson, B.A., Fernandes, H.M., Murray, G.D.,
Teasdale, G.M., Hope, D.T., Karimi, A., Shaw, M.D., and Barer, D.H.
(2005). Early surgery versus initial conservative treatment in patients
with spontaneous supratentorial intracerebral haematomas in the In-
ternational Surgical Trial in Intracerebral Haemorrhage (STICH): a
randomised trial. Lancet 365, 387397.
19. Drummond, M., Barbieri, M., Cook, J., Glick, H.A., Lis, J., Malik, F.,
Reed, S.D., Rutten, F., Sculpher, M., and Severens, J. (2009). Trans-
ferability of economic evaluations across jurisdictions: ISPOR Good
Research Practices Task Force report. Value Health 12, 409418.
20. Manca, A., Sculpher, M.J., and Goeree, R. (2010). The analysis of
multinational cost-effectiveness data for reimbursement decisions: a
critical appraisal of recent methodological developments. Pharma-
coeconomics 28, 10791096.
21. Campbell and Cochrane Economics Methods Group and the Evidence
for Policy and Practice Information and Co-ordinating Centre (CCEMG-
EPPI). CCEMG-EPPI Centre Cost Converter (V.1.2).
22. World Bank. GNI Per Capita. How we classify countries World Bank.
http://data.worldbank.org/about/country-classifications (last accessed
March 20, 2014).
23. Mendelow, A.D., Gregson, B.A., Rowan, E.N., Murray, G.D.,
Gholkar, A., Mitchell, P.M., and Investigators, S.I. (2013). Early
surgery versus initial conservative treatment in patients with sponta-
neous supratentorial lobar intracerebral haematomas (STICH II): a
randomised trial. Lancet 382, 397408.
24. Francis , R., Gregson, B., Mendelow, A., and Rowan, E. (2015).
Characteristics of traumatic intracerebral haemorrhage: an assessment
of screening logs from the STITCH(Trauma) Trial. (In press.)
25. Francis, R., Gregson, B.A., and Mendelow, A.D. (2014). Attitudes to
intracranial pressure monitoring of traumatic intracerebral haemor-
rhage. Br. J. Neurosurg. 28, 663665.
Address correspondence to:
Barbara A. Gregson, PhD
Neurosurgical Trials Group
Wolfson Research Center
Campus for Ageing and Vitality
Newcastle upon Tyne NE4 5PL
United Kingdom
E-mail: Barbara.Gregson@ncl.ac.uk
(Appendix follows/)
 SURGERY FOR TRAUMATIC ICH RCT: STITCH(TRAUMA) 1323

Appendix

STITCH (TRAUMA) Investigators India: Bangalore, BGS Global Hospital (1): S.A.V. Rao; N.K.
Venkataramanaa; Bangalore, NIMHANS (5): S. Somanna;
Principal investigators: A. David Mendelow, MB, BCh, PhD,
K.V.L.N. Rao; J. lal Gangadharan; Calcutta, AMRI Hospitals
FRCS; Barbara A. Gregson, BSc, PhD, FSS; Patrick M. Mitchell,
(0): R.N. Bhattacharya; Chennai, Fortis Malar Hospital (1): K.
BA, MB, BChir, BSc, FRCS, PhD; Andy Unterberg, MD, PhD;
Sridhar; G. Venkatprasanna; Dehradun, Himalayan Institute
Elaine M. McColl, BA, MSc, PhD; Iain R. Chambers, BSc, PhD,
of Medical Sciences (13): K.K. Bansal; C. Gupta, R. Kumar;
CEng, FIPEM; Paul McNamee, MA, MSc, PhD.
Lucknow, King Georges Medical University (erstwhile CSM
Steering committee: Mr. J. Steers (independent chairman); Mr.
Medical University) (29): S.K. Singh; C. Srivastava; B.K. Ojha;
Andy Vail (statistician); Dr. D. Birchall (neuroradiologist, inde-
A. Chandra; Ludhiana, Christian Medical College & Hospital
pendent member); Mr. Jake Timothy (neurosurgeon, independent
(3): S.S. Grewal; B. Gupta; Maharashtra, Acharya Vinoba
member); Professor Luke Vale (health economist, independent
Bhave Rural Hospital (3): A. Agrawal; Mullana (Ambala), MM
member); Mr. A. White (lay member, Headway); Mr. D. OMeara
Downloaded from online.liebertpub.com by 36.80.211.182 on 09/24/17. For personal use only.

Institute of Medical Sciences and Research (1): A. Agrawal;

(lay member, UKABIF); Professor A.D. Mendelow; Dr. B.A.
Mysore, Mysore Clinisearch (2): A. Sangli; New Delhi, All
Gregson; Mr. P.M. Mitchell; Dr. A. Unterberg; Professor E.M.
India Institute of Medical Sciences (8): P. Sarat Chandra; B.S.
McColl; Dr. I.R. Chambers; Professor P. McNamee; Dr. E.N.
Sharma; Visakhapatnam, Care Hospital (8): P.V. Ramana; P.M.
Rowan; Dr. D. Boyers; Dr. R. Francis.
Jagannath.
Data monitoring committee: Professor P. Hutchinson (chairman
Latvia: Riga, Pauls Stradins Clinical University Hospital (0): E.
and neurosurgeon); Professor G.D. Murray (independent statisti-
Valeinis.
cian); Dr. A. Gholkar (neuroradiologist).
Lithuania: Kaunas, Kaunas University of Health Sciences Hos-
Trial management team: Professor A. David Mendelow (chief
pital (2): A. Tamasauskas, R. Vilcinis; Klaipeda, Klaipeda Uni-
investigator); Dr. Barbara A. Gregson (principal research associate
Journal of Neurotrauma 2015.32:1312-1323.

versity Hospital (0): A. Gvazdaitis.

and trial director); Dr. Elise Rowan (senior research associate and
trial manager); Dr. Richard Francis (research associate and data
manager); Dr. Dwayne Boyers (research associate and economist);
Miss Helen Atkinson (trial administrator 20092010); Miss Cour-
tenay Howe (trial administrator 20102013); Mr. Patrick Mitchell
(neurosurgeon).
Radiological committee: A. Hassani; Y.K. Yap; L. Yap; A. Gholkar.
Writing committee: A.D. Mendelow; B.A. Gregson; E.N. Ro-
wan; R. Francis; E. McColl; P. McNamee; I.R. Chambers; A.
Unterberg; D. Boyers; P.M. Mitchell.
List of Center Investigators by Country and Center
Together with the Number of Patients Recruited
(In this list, centers that recorded zero patients
did return screening information)
Canada: Toronto, St Michaels Hospital (0): R.L. Macdonald.
Bulgaria: Sofia, University Hospital Pirogov (2): N. Gabrovsky;
N. Velinov.
Czech Republic: Brno, University Hospital Brno (1): M.
Smrcka; G. Hanoun.
Egypt: Alexandria, Alexandria University Hospitals (3): O.S.
Abdelaziz; I.H. Zidan.
Germany: Heidelberg, Heidelberg University Hospital (2): A.W.
Unterberg; C. Beynon; Munich, Bogenhausen Academic Teaching
Hospital, Technical University of Munich (1): C.B. Lumenta; D.B.
Schul; Ulm, University of Ulm School of Medicine (0): M.-E.
Halatsch; A. Pala.
Hungary: Szeged, University of Szeged.
Neurosurgery (0): P. Barzo; B. Fulop.
Malaysia: Malaysia, Johor Bahru, Hospital Sultanah Aminah
Johor Bahru (1): N.A.A. Rahman A. Ali; Kota Bharu, Kelantan,
Universiti Sains Malaysia (6): J.M. Abdullah; T.Y. Chin.
Nepal: Biratnagar, Neuro Hospital (16): Y.B. Roka; P.R. Puri.
Pakistan: Peshawar, Northwest General Hospital & Research
Center Peshawar (2): T. Khan; F. Filza.
Peoples Republic of China: Beijing, Tiantan Hospital affili-
ated to Capital Medical University (30): J. Zhao; L. Xu; J. Li;
Shanghai, Huashan Hospital, Fudan University (9): Y. Sun; J.
Hu; Tianjin, Tianjin Medical University General Hospital (4):
S. Yang; R. Jiang.
Romania: Cluj-Napoca, Cluj County Emergency Hospital (1):
I.S. Florian; M. Rus; Timisoara, Emergency County Hospital Ti-
misoara (7): H. Ples; S.M. Marius.
Spain: Santander, University Hospital Marques de Valdecilla
(2): A. Vazquez-Barquero; Valladolid, Universitario Ro Hortega
(1): R. Sarabia; I. Arrese.
United Kingdom: Cambridge, Addenbrookes Hospital, Cam-
bridge University Hospitals NHS Foundation Trust (0): P.J. Kirk-
patrick; A.G. Kolias; Dundee, Ninewells Hospital and Medical
School (2): S. Eljamel; Haywards Heath, Hurstwood Park Neu-
roscience Centre (0): G. Critchley; J. Norris; Newcastle, Royal
Victoria Infirmary (3): P. Bhattathiri; N. Ross; Southampton,
Southampton General Hospital (1): A. Belli; D. Bulters.
United States of America: Los Angeles, Los Angeles County &
University of Southern California Medical Center (0): J.P. Gruen;
Philadelphia, Temple University Hospital (0): M.W. Weaver; F.
Sultan; Portland, Legacy Emanuel Medical Center (0): J.W. Chen;
S. Staat.